"""
CS101 example of scraping a webpage for email addresses
"""

import urllib.request

def get_emails(URL):
    """
    Get the email addresses from the given webpage by extracting the strings
    following 'mailto:' until the next double-quote character
    and return them in a list
    """
    webpage = urllib.request.urlopen(URL)
    search_string = "mailto:"
    emails = []
    
    for line in webpage:        
        line = line.decode('utf-8')
        if search_string in line:
            begin_index = line.find(search_string) + len(search_string)
            end_index = line.find('"', begin_index)
            emails.append(line[begin_index:end_index])
    
    return emails

# sample call to get_emails() sending in a sample URL:
# get_emails("http://www.middlebury.edu/academics/cs/faculty/")